Pseudowires in VPLS bridge domains

Key features of VFIs in VPLS

Virtual forwarding instances (VFIs) are virtual bridge ports used in VPLS networks to interconnect the pseudowire mesh for each VPLS instance. VFIs perform native bridging functions required for packet forwarding, including learning and aging of MAC addresses, and directing data based on destination MAC addresses.

• Provide virtual bridge ports within the VPLS instance.

• Perform native bridging functions such as forwarding based on destination MAC addresses, source MAC address learning, and aging.

• Enable PE routers to make packet-forwarding decisions by referencing the VFI associated with a particular VPLS instance.

• Support connection of multiple attachment circuits belonging to a given VPLS.

Here is the sample configuration for a VFI under the bridge domain.

Router# configure
Router(config)# l2vpn
Router(config-l2vpn)# bridge group BG1
Router(config-l2vpn-bg)# bridge-domain BD1
Router(config-l2vpn-bg-bd)# vfi V1
Router(config-l2vpn-bg-bd-vfi)# neighbor 10.0.0.2 pw-id 1
Router(config-l2vpn-bg-bd-vfi-pw)# commit

The corresponding running configuration is:

l2vpn
 bridge group BG1
  bridge-domain BD1
   vfi V1
    neighbor 10.0.0.2 pw-id 1
  !
!

VPLS pseudowires support several transport types, enabling flexible and scalable network designs for multipoint Layer 2 VPN services. Each transport type provides different underlying mechanisms for establishing connectivity across provider networks.

The supported transport types for VPLS pseudowires are:

• LDP: Utilizes LDP for signaling and establishing pseudowire tunnels.

• Segment routing: Leverages segment routing for simplified path management and traffic engineering.

• LDP over TE: Combines LDP signaling with TE tunnels for enhanced path control.

• BGP-LU for inter-AS C topology: Uses BGP with Label Unicast to support inter-AS connections in C topology scenarios.

Ethernet port mode (Type 5)

These pseudowire types are supported for MPLS:

• Both ends of the pseudowire connect to Ethernet ports and transport a complete ethernet trunk.

• Frames received on a main interface or subinterface are transported by the ingress PE.

• This mode eliminates the need for dummy VLAN tags and reduces overhead.

• Frame tagging is no longer necessary.

• The ingress PE does not remove incoming VLAN tags; they are transported over the pseudowire.

• Type 4 inserts an extra dummy tag with VLAN 0 onto the frame, which is removed at the remote end.

• Allows service providers to segregate traffic per customer using VLANs.

• Each VLAN on a customer-to-provider link can be configured as a separate Layer 2 VPN connection.

Key considerations:

• VLAN-based VC type 4 pseudowires transport VLAN tags by pushing a dummy tag at the ingress when only a single tag is present.

• If multiple VLAN tags are pushed, a dummy tag is unnecessary—they are transported directly.

• The dummy tag (if added) is removed on the remote router before egress.

• Mixing Type 4 and Type 5 pseudowires in the same Virtual Forwarding Instance (VFI) is not supported: all pseudowires in a VFI must be of the same type.

The pseudowire control-word serves several key functions in AToM packet transport:

• Padding small packets: If an AToM packet is below the minimum length, the frame is padded to meet Ethernet requirements. The control-word includes a length field that helps the egress PE router determine whether padding was added and remove it before forwarding.

• Carrying control bits: The control-word transports control bits from the Layer 2 header of the protocol being carried.

• Sequencing transported frames: The control-word preserves the sequence of frames. Each packet sent over the pseudowire receives an incrementing sequence number, starting with 1 and continuing to 65535. Out-of-sequence packets are dropped; packets are never re-ordered.

• Load balancing: The control-word enables correct identification of Ethernet PW packets and prevents misordering caused by equal-cost multipath (ECMP) path selection. The control-word, inserted after the MPLS label, separates payload from MPLS and carries Layer 2 control bits and supports sequencing.

• Fragmentation signaling:

  The control-word indicates payload fragmentation status using two bits:

  • 00: Unfragmented

  • 01: First fragment

  • 10: Last fragment

  • 11: Intermediate fragment

Flow-aware transport pseudowires (FAT PW) enhance load balancing for traffic carried over pseudowires. Routers typically load balance traffic using the lowermost label in the label stack. In standard pseudowires, this label remains the same for all flows, which can cause uneven traffic distribution and asymmetric load balancing.

Key characteristics of flow-aware transport pseudowires:

• FAT PW identifies individual packet flows within a pseudowire.

• FAT PW creates a flow label based on each packet flow entering the pseudowire.

• FAT PW inserts the flow label as the lowermost label in the packet.

• Routers use the flow label to make load-balancing decisions.

FAT PW enables routers to identify individual flows within a pseudowire and leverage that flow information for more effective load balancing. This capability is especially valuable in core networks using equal-cost multipaths (ECMP).